Approximation solution of boundary values of partial differential equations by using artificial neural networks

Abstract EN

This paper contains aspects of the study of approximation solution of boundary values of partial differential equations using artificial neural networks.

This study present an explanation of ridge functions .

Then, various motivations of their study are given.

The researcher present an algorithm in order to achieve the intended approximation by using backpropagation neural network.

To assess has been the performance of the discussed method, a comparison of the finite difference method has been carried out so that the ability and robustness of the presented method becomes clear.